The present invention relates generally to searching and classifying unstructured information about geographic objects within a bounded area of an electronic map.
Technology for performing a text based search of a geographic area based on information stored in a point of interest database is generally known. Generally, information about points of interest is stored in a structured database including the point of interest location and characteristics. For example, restaurants are stored as points of interest in a structured database containing fields describing restaurant properties (e.g. type of food, location, price, etc.). A search with certain descriptive characteristics (e.g., pizza) within a zip code displays a list of matching restaurants, and each matching restaurant is represented by a symbol and located on the electronic map.
More recently, unstructured items of information (e.g., blogs, articles, documents, images, text, media, etc.) are geotagged and posted to the Internet. Geotagging is the process of adding geographical identification metadata to content. Recently, as the result of the proliferation of global positioning technology, images posted to the Internet often have geotags. The geo-tagged content may be relevant and useful for geographic object searching.
Unstructured information, without any metadata suggesting a category for classification, may be classified with statistical methods, machine learning methods, or ontological methods. An ontology defines the terms used to describe and represent an area of knowledge. Information may be classified using an ontology domain schema containing domain entities organized into categories and interconnected by relationships.